Statistical Prediction of Product Quality in Batch Processes with Limited Batch-Cycle Data

Zhiqiang Ge,Zhihuan Song,Furong Gao

DOI: https://doi.org/10.1021/ie202554r

2012-01-01

Industrial & Engineering Chemistry Research

Abstract:This paper develops a modeling approach to address the end-of-batch product quality prediction problem for batch processes with limited batch-cycle data. Generally, those batch processes that have multiple phases are the focus of the present paper. Different from the traditional multiway/phase-based partial least-squares (PLS) method, which unfolds the three-way data set through the batch direction, the proposed method unfolds the data set through the variable direction, in order to generate more training data samples. Reproducing the product quality data with the noise injection method allows a statistical model to be developed in each phase of the batch process. This, however, does not remove the nonlinearity of the batch process data, as practically addressed by the typical batch normalization. Therefore, a nonlinear regression model is subsequently introduced to handle this problem for product quality prediction modeling. To compare the performance of linear and nonlinear statistical models, phase-based PLS and relevance vector machine models have both been developed for prediction of product quality in an industrial injection molding process.